Helenium Virginicum (Asteraceae): Antiherbivore Defense?
Total Page:16
File Type:pdf, Size:1020Kb
Banisteria, Number 13, 1999 C 1999 by the Virginia Natural History Society Factors Controlling the Distribution and Abundance of the Narrow Endemic, Helenium virginicum (Asteraceae): Antiherbivore Defense? John S. Knox, Frank W. Steams Jr., & Charles K. Dietzel, Department of Biology Washington and Lee University Lexington, VA 24450 INTRODUCTION to increase with increasing stochasticity of the environ- ment and increasing variability of population size (Good- Helenium virginicum Blake, the Virginia sneezeweed, man, 1987; Pimm et al., 1988; Menges, 1991). It is not is a perennial herbaceous member of the Asteraceae that surprising then, that the Kennedy Mountain Meadow is globally endemic to about 22 sinkhole ponds in Augusta population has twice crashed after unusually lengthy and Rockingham Counties, Virginia (Knox, 1995; Van periods of inundation that lasted- for 16 and 20 months, in Aistine, 1996). The state of Virginia lists the species as 1989 — 90, and 1995 - 97, respectively. Each time that Endangered (Porter & Wieboldt, 1991), and the U.S. Fish extinction of extant plants has occurred, recruitment from and Wildlife Service has recently listed the plant as a a seed bank regenerated the population within one month Threatened species under the U.S. Endangered Species of drawdown. In each case, local extinction was associ- Act (U.S. Fish and Wildlife Service, 1998). ated with heavy growth of floating aquatic vegetation. The narrow endemism of Helenium virginicum seems Evidence suggests that these local extinctions might be to reflect its limitation to rare sites where competition explained in part by the shade intolerance of H. virgini- with other species has been reduced by an unusual cum. Knox (1997) found that distribution of H. virginicum combination of stressful edaphic and hydrologic condi- plants within sinkhole basins reflected shade intolerance tions (Knox, 1997). The clay soils at the sinkhole ponds of the species. At sites surrounded by tall forest with a where this species grows have a low pH, averaging 4.5, relatively dense canopy, plants were absent from the south low levels of B, Ca, K, Mg, and P, and high levels of Al side' of the basins where shade is greatest. Knox (1997) and As. The combination of low pH and high Al is the observed that deeper sinkhole ponds tended to have H. most important edaphic condition that limits the agricul- virginicum plants distributed in a belt, beyond the shade tural use of acid soils worldwide (Barinaga, 1997; Fuente of dense vegetation near the shoreline, but not too deep to et al., 1997), and is well known to impair uptake of the experience .too much shade from the water column during required macronutrients Ca, Mg, and K (Foy, 1974; the winter period of inundation. Knox (1997) found that Taylor, 1988). the patterns of natural recruitment and the survival / of Growing conditions in the sinkhole ponds inhabited by transplants at Kennedy Mountain M&adow also suggested H. virginicum are made more stressful by dramatic that H. virginicum is shade intolerant. Also, viable seeds seasonal changes in water depth (Fig. 1), with months of that have broken dormancy do not germinate under continuous flooding in cooler seasons alternating with favorable temperatures when the seeds are in the dark or periods of drawdown, when the basins lack any standing under a standing column of water that contains green water, during warmer times of the year. Year-to-year plants (J. J. W. Harvey and J. S. Knox, unpublished data). variation in the duration of flooding may be great (Fig. 1); The 'plant grows year round, as is evident from the more long periods of inundation have been associated with narrow leaves produced during winter inundation (J. S. precipitous reductions in one population of H. virginicum Knox, personal observation). We hypothesize that local (Fig. 2) intensively studied for 12 years at Kennedy extinctions occurred after long-term inundation because Mountain Meadow (Knox, 1997; J. S. Knox, unpublished H. virginicum plants growing on the bottom of the basin data). In other species, risk of extinction has been found were shaded below their light compensation point by 96 BANISTERIA NO. 13, 1999 11 111111111111111 X86 1987 1988 75-- 50- 25- I It 11 ND ND 11 J D J D J 1992 1993 1994 1995 Date (Year, Month) Fig. I. Water depth (cm) measured monthly at the center (nearly the deepest point) at Kennedy Mountain. Meadow. D is December; J is June; ND indicates no data collected. floating aquatic vegetation. extinction in cases where population size falls to fewer Another factor that may make H. virginicum vulnera- than 50 flowering individuals. We think that for H. ble to local extinction, and thus contribute to its narrow virginicum, the high year-to-year variability in hydro- endemism, is the plant's sporophytic self-incompatible period, with its dramatic impact on population size, in breeding system (Messmore & Knox, 1,997). Successful conjunction with self-incompatibility, has created a seed production requires pollen with a different mating situation in which the risk of local extinction is high and phenotype than is present on the stigma. Computer the probability of establishing a new population is low. simulations (Byers & Meagher, 1992), as well as field The largest and densest populations of H virginicum studies (DeMauro, 1993) with other species, have re- occur at sites that have been disturbed by human activities vealed that this mating system increases the risk of local such as mowing, grazing, and cutting trees. During a KNOX, STEARNS, & DIETZEL: ANTIHERBIVORE DEFENSE 97 period of population expansion in 1987, Knox (unpub- other species of Helenium are known to be toxic and lished data) found densities of over 400 plants/m2 at one unpalatable to vertebrate (Hesker, 1982; Anderson et al., sinkhole pond that had been disturbed by mowing, in 1983) and invertebrate (Amason et al., 1987) animals. comparison to densities up to 83 plants/m 2 at an especially The toxicity of Helenium (Anderson et al., 1986; Arnason dense area of an undisturbed site that same year., Knox et al., 1987) and other plant genera (Burnett et al., 1977) (1997) speculated that the factors responsible for the has been associated with sesquiterpene lactones. Since H. Much greater densities of H. virginicurn at disturbed sites virginicum is known to contain a specific sesquiterpene may have been activities that reduced the cover of com- lactone, Virginolide, (Herz & Santhanam, 1967), and peting plants, by trees having been cut, and the sites handling the plant leaves a bitter residue on the hands (J. having been mowed and grazed. Knox (1997) called for Knox, personal observation), we suspect that H. virgini- future study of the impact of tree cutting, mowing, and cum may be unpalatable to generalist herbivores. grazing on H virginicum, suggesting that this information To test this hypothesis, we established a common could be useful for management of the species. garden study site where native vertebrate herbivores were We suspect that H. virginicum may be unpalatable to known to be common, and planted large numbers of H. herbivores; that grazing herbivores may selectively reduce virginicum in close association with large numbers of an the biomass of co-occurring plant species, thus reducing attractant plant species palatable to those herbivores. We competition for light between the rare plant and its sought evidence of selective avoidance of H. virginicum •associates. This hypothesis seems plausible given that by herbivorous animals that fed on adjacent attractant ( 25000- I- < 20000- C:3 < 2 cm W 2 cm 15000- - Flowering Fa- O 10000- U. uj Z Z 5000- < 0 d ' — 1986 1987 1988 1989 1 '690 1991 1992 1993 1994 1995 1996 1997 YEAR Fig. 2. Total number of Helenium virginicum plants growing within the 1 ha basin of Kennedy Mountain Meadow by stage class, censused in the fall of each year. Flowering refers to plants that flowered that season; < 2 cm refers to the length of the longest rosette leaf on plants that did not flower; > 2 cm refers to the longest rosette leaf on plants that did not flower. Numbers are estimates made by extrapolating from annual census data taken at 53 permanent quadrats, except for 1994 and 1997, when total counts of all plants in the basin were made. 98 BANISTERIA NO. 13, 1999 plants. We reasoned that if substantial herbivOry of the senescence on 5 June 1998, a second cohort of beans was attractant plant occurred, while interspersed with large started from seeds, by planting three bean seeds around numbers of H. virginicum, and the rare plant experienced each adult bean plant. On 17 June 1998, all first cohort little or no herbivory, we could infer that H virginicum is bean plants were removed and the second cohort bean unpalatable to herbiVores. We were also interested to seedlings was thinned \to leave one seedling where each learn if H. virginicum would reduce the extent of herb- first cohort bean plant had been. The treatment quadrats ivory on the attractant plant by comparing herbivory of were weeded. We inspected the experimental plants nearly attractant plants when grown in monoculture with those every day, but observed no evidence of herbivory until grown interspersed with H. virginicum. Such repellent after the herbaceous vegetation in the surrounding old plants have been described in other taxa (Atsatt & field was mowed on 6 July. On 8 July 1998, we began to O'Dowd, 1976). '‘ observe substantial evidence of herbivory, and that evidence increased until 23 July, when we ended the MATERIALS AND METHODS experiment by harvesting the above ground biomass of each bean plant by cutting the stem at soil level. Each In May 1998, we established a common garden in the plant was placed separately in an aluminum pan and dried center of an old field of approximately 0.5 ha on the in an oven for three days at 95 C.